Should a Bot be Sarcastic? Understanding User Preferences Towards Sarcasm Generation

Anonymous

Abstract: Previous sarcasm generation research has focused on how to generate text that people perceive as sarcastic to create more human-like interactions. In this paper, we argue that we should first turn our attention to the question of when sarcasm should be generated, finding that human annotators consider many inputs to be unfit for sarcastic responses. Next, we introduce a theory-driven framework for sarcasm generation which allows us to better control the linguistic devices used during the generation process in order to measure their impact on sarcasm perception, finding that pragmatic insincerity and emotional markers are crucial elements in generating recognizable sarcasm.